Complexity and Algorithms for Nonlinear Optimization Problems
Total Page:16
File Type:pdf, Size:1020Kb
Load more
Recommended publications
-
Nonlinear Optimization Using the Generalized Reduced Gradient Method Revue Française D’Automatique, Informatique, Recherche Opé- Rationnelle
REVUE FRANÇAISE D’AUTOMATIQUE, INFORMATIQUE, RECHERCHE OPÉRATIONNELLE.RECHERCHE OPÉRATIONNELLE LEON S. LASDON RICHARD L. FOX MARGERY W. RATNER Nonlinear optimization using the generalized reduced gradient method Revue française d’automatique, informatique, recherche opé- rationnelle. Recherche opérationnelle, tome 8, no V3 (1974), p. 73-103 <http://www.numdam.org/item?id=RO_1974__8_3_73_0> © AFCET, 1974, tous droits réservés. L’accès aux archives de la revue « Revue française d’automatique, in- formatique, recherche opérationnelle. Recherche opérationnelle » implique l’accord avec les conditions générales d’utilisation (http://www.numdam.org/ conditions). Toute utilisation commerciale ou impression systématique est constitutive d’une infraction pénale. Toute copie ou impression de ce fi- chier doit contenir la présente mention de copyright. Article numérisé dans le cadre du programme Numérisation de documents anciens mathématiques http://www.numdam.org/ R.A.LR.O. (8* année, novembre 1974, V-3, p. 73 à 104) NONLINEAR OPTIMIZATION USING THE-GENERALIZED REDUCED GRADIENT METHOD (*) by Léon S. LÀSDON, Richard L. Fox and Margery W. RATNER Abstract. — This paper describes the principles and logic o f a System of computer programs for solving nonlinear optimization problems using a Generalized Reduced Gradient Algorithm, The work is based on earlier work of Âbadie (2). Since this paper was written, many changes have been made in the logic, and significant computational expérience has been obtained* We hope to report on this in a later paper. 1. INTRODUCTION Generalized Reduced Gradient methods are algorithms for solving non- linear programs of gênerai structure. This paper discusses the basic principles of GRG, and constructs a spécifie GRG algorithm. The logic of a computer program implementing this algorithm is presented by means of flow charts and discussion. -
Combinatorial Structures in Nonlinear Programming
Combinatorial Structures in Nonlinear Programming Stefan Scholtes¤ April 2002 Abstract Non-smoothness and non-convexity in optimization problems often arise because a combinatorial structure is imposed on smooth or convex data. The combinatorial aspect can be explicit, e.g. through the use of ”max”, ”min”, or ”if” statements in a model, or implicit as in the case of bilevel optimization where the combinatorial structure arises from the possible choices of active constraints in the lower level problem. In analyzing such problems, it is desirable to decouple the combinatorial from the nonlinear aspect and deal with them separately. This paper suggests a problem formulation which explicitly decouples the two aspects. We show that such combinatorial nonlinear programs, despite their inherent non-convexity, allow for a convex first order local optimality condition which is generic and tight. The stationarity condition can be phrased in terms of Lagrange multipliers which allows an extension of the popular sequential quadratic programming (SQP) approach to solve these problems. We show that the favorable local convergence properties of SQP are retained in this setting. The computational effectiveness of the method depends on our ability to solve the subproblems efficiently which, in turn, depends on the representation of the governing combinatorial structure. We illustrate the potential of the approach by applying it to optimization problems with max-min constraints which arise, for example, in robust optimization. 1 Introduction Nonlinear programming is nowadays regarded as a mature field. A combination of important algorithmic developments and increased computing power over the past decades have advanced the field to a stage where the majority of practical prob- lems can be solved efficiently by commercial software. -
Nonlinear Integer Programming ∗
Nonlinear Integer Programming ∗ Raymond Hemmecke, Matthias Koppe,¨ Jon Lee and Robert Weismantel Abstract. Research efforts of the past fifty years have led to a development of linear integer programming as a mature discipline of mathematical optimization. Such a level of maturity has not been reached when one considers nonlinear systems subject to integrality requirements for the variables. This chapter is dedicated to this topic. The primary goal is a study of a simple version of general nonlinear integer problems, where all constraints are still linear. Our focus is on the computational complexity of the problem, which varies significantly with the type of nonlinear objective function in combination with the underlying combinatorial structure. Nu- merous boundary cases of complexity emerge, which sometimes surprisingly lead even to polynomial time algorithms. We also cover recent successful approaches for more general classes of problems. Though no positive theoretical efficiency results are available, nor are they likely to ever be available, these seem to be the currently most successful and interesting approaches for solving practical problems. It is our belief that the study of algorithms motivated by theoretical considera- tions and those motivated by our desire to solve practical instances should and do inform one another. So it is with this viewpoint that we present the subject, and it is in this direction that we hope to spark further research. Raymond Hemmecke Otto-von-Guericke-Universitat¨ Magdeburg, FMA/IMO, Universitatsplatz¨ 2, 39106 Magdeburg, Germany, e-mail: [email protected] Matthias Koppe¨ University of California, Davis, Dept. of Mathematics, One Shields Avenue, Davis, CA, 95616, USA, e-mail: [email protected] Jon Lee IBM T.J. -
An Improved Sequential Quadratic Programming Method Based on Positive Constraint Sets, Chemical Engineering Transactions, 81, 157-162 DOI:10.3303/CET2081027
157 A publication of CHEMICAL ENGINEERING TRANSACTIONS VOL. 81, 2020 The Italian Association of Chemical Engineering Online at www.cetjournal.it Guest Editors: Petar S. Varbanov, Qiuwang Wang, Min Zeng, Panos Seferlis, Ting Ma, Jiří J. Klemeš Copyright © 2020, AIDIC Servizi S.r.l. DOI: 10.3303/CET2081027 ISBN 978-88-95608-79-2; ISSN 2283-9216 An Improved Sequential Quadratic Programming Method Based on Positive Constraint Sets Li Xiaa,*, Jianyang Linga, Rongshan Bia, Wenying Zhaob, Xiaorong Caob, Shuguang Xianga a State Key Laboratory Base for Eco-Chemical Engineering, College of Chemical Engineering, Qingdao University of Science and Technology, Zhengzhou Road No. 53, Qingdao 266042, China b Chemistry and Chemistry Engineering Faculty,Qilu Normal University, jinan, 250013, Shandong, China [email protected] SQP (sequential quadratic programming) is an effective method to solve nonlinear constraint problems, widely used in chemical process simulation optimization. At present, most optimization methods in general chemical process simulation software have the problems of slow calculation and poor convergence. To solve these problems , an improved SQP optimization method based on positive constraint sets was developed. The new optimization method was used in the general chemical process simulation software named optimization engineers , adopting the traditional Armijo type step rule, L-BFGS (Limited-Memory BFGS) algorithm and rules of positive definite matrix. Compared to the traditional SQP method, the SQP optimization method based on positive constraint sets simplifies the constraint number of corresponding subproblems. L-BFGS algorithm and rules of positive definite matrix simplify the solution of the second derivative matrix, and reduce the amount of storage in the calculation process. -
Appendix a Solving Systems of Nonlinear Equations
Appendix A Solving Systems of Nonlinear Equations Chapter 4 of this book describes and analyzes the power flow problem. In its ac version, this problem is a system of nonlinear equations. This appendix describes the most common method for solving a system of nonlinear equations, namely, the Newton-Raphson method. This is an iterative method that uses initial values for the unknowns and, then, at each iteration, updates these values until no change occurs in two consecutive iterations. For the sake of clarity, we first describe the working of this method for the case of just one nonlinear equation with one unknown. Then, the general case of n nonlinear equations and n unknowns is considered. We also explain how to directly solve systems of nonlinear equations using appropriate software. A.1 Newton-Raphson Algorithm The Newton-Raphson algorithm is described in this section. A.1.1 One Unknown Consider a nonlinear function f .x/ W R ! R. We aim at finding a value of x so that: f .x/ D 0: (A.1) .0/ To do so, we first consider a given value of x, e.g., x . In general, we have that f x.0/ ¤ 0. Thus, it is necessary to find x.0/ so that f x.0/ C x.0/ D 0. © Springer International Publishing AG 2018 271 A.J. Conejo, L. Baringo, Power System Operations, Power Electronics and Power Systems, https://doi.org/10.1007/978-3-319-69407-8 272 A Solving Systems of Nonlinear Equations Using Taylor series, we can express f x.0/ C x.0/ as: Â Ã.0/ 2 Â Ã.0/ df .x/ x.0/ d2f .x/ f x.0/ C x.0/ D f x.0/ C x.0/ C C ::: dx 2 dx2 (A.2) .0/ Considering only the first two terms in Eq. -
A Sequential Quadratic Programming Algorithm with an Additional Equality Constrained Phase
A Sequential Quadratic Programming Algorithm with an Additional Equality Constrained Phase Jos´eLuis Morales∗ Jorge Nocedal † Yuchen Wu† December 28, 2008 Abstract A sequential quadratic programming (SQP) method is presented that aims to over- come some of the drawbacks of contemporary SQP methods. It avoids the difficulties associated with indefinite quadratic programming subproblems by defining this sub- problem to be always convex. The novel feature of the approach is the addition of an equality constrained phase that promotes fast convergence and improves performance in the presence of ill conditioning. This equality constrained phase uses exact second order information and can be implemented using either a direct solve or an iterative method. The paper studies the global and local convergence properties of the new algorithm and presents a set of numerical experiments to illustrate its practical performance. 1 Introduction Sequential quadratic programming (SQP) methods are very effective techniques for solv- ing small, medium-size and certain classes of large-scale nonlinear programming problems. They are often preferable to interior-point methods when a sequence of related problems must be solved (as in branch and bound methods) and more generally, when a good estimate of the solution is available. Some SQP methods employ convex quadratic programming sub- problems for the step computation (typically using quasi-Newton Hessian approximations) while other variants define the Hessian of the SQP model using second derivative informa- tion, which can lead to nonconvex quadratic subproblems; see [27, 1] for surveys on SQP methods. ∗Departamento de Matem´aticas, Instituto Tecnol´ogico Aut´onomode M´exico, M´exico. This author was supported by Asociaci´onMexicana de Cultura AC and CONACyT-NSF grant J110.388/2006. -
GRASP for NONLINEAR OPTIMIZATION in This Paper, We
GRASP FOR NONLINEAR OPTIMIZATION C.N. MENESES, P.M. PARDALOS, AND M.G.C. RESENDE ABSTRACT. We propose a Greedy Randomized Adaptive Search Procedure (GRASP) for solving continuous global optimization problems subject to box constraints. The method was tested on benchmark functions and the computational results show that our approach was able to find in a few seconds optimal solutions for all tested functions despite not using any gradient information about the function being tested. Most metaheuristcs found in the literature have not been capable of finding optimal solutions to the same collection of functions. 1. INTRODUCTION In this paper, we consider the global optimization problem: Find a point x∗ such that n f (x∗) f (x) for all x X, where X is a convex set defined by box constraints in R . For ≤ 2 instance, minimize f (x ;x ) = 100(x x2)2 +(1 x )2 such that 2 x 2 for i = 1;2. 1 2 2 − 1 − 1 − ≤ i ≤ Many optimization methods have been proposed to tackle continuous global optimiza- tion problems with box constraints. Some of these methods use information about the function being examined, e.g. the Hessian matrix or gradient vector. For some functions, however, this type of information can be difficult to obtain. In this paper, we pursue a heuristic approach to global optimization. Even though several heuristics have been designed and implemented for the problem discussed in this paper (see e.g. [1, 15, 16, 17]), we feel that there is room for yet an- other one because these heuristics were not capable of finding optimal solutions to several functions in standard benchmarks for continuous global optimization problems with box constraints. -
Parallel Technique for the Metaheuristic Algorithms Using Devoted Local Search and Manipulating the Solutions Space
applied sciences Article Parallel Technique for the Metaheuristic Algorithms Using Devoted Local Search and Manipulating the Solutions Space Dawid Połap 1,* ID , Karolina K˛esik 1, Marcin Wo´zniak 1 ID and Robertas Damaševiˇcius 2 ID 1 Institute of Mathematics, Silesian University of Technology, Kaszubska 23, 44-100 Gliwice, Poland; [email protected] (K.K.); [email protected] (M.W.) 2 Department of Software Engineering, Kaunas University of Technology, Studentu 50, LT-51368, Kaunas, Lithuania; [email protected] * Correspondence: [email protected] Received: 16 December 2017; Accepted: 13 February 2018 ; Published: 16 February 2018 Abstract: The increasing exploration of alternative methods for solving optimization problems causes that parallelization and modification of the existing algorithms are necessary. Obtaining the right solution using the meta-heuristic algorithm may require long operating time or a large number of iterations or individuals in a population. The higher the number, the longer the operation time. In order to minimize not only the time, but also the value of the parameters we suggest three proposition to increase the efficiency of classical methods. The first one is to use the method of searching through the neighborhood in order to minimize the solution space exploration. Moreover, task distribution between threads and CPU cores can affect the speed of the algorithm and therefore make it work more efficiently. The second proposition involves manipulating the solutions space to minimize the number of calculations. In addition, the third proposition is the combination of the previous two. All propositions has been described, tested and analyzed due to the use of various test functions. -
A New Successive Linearization Approach for Solving Nonlinear Programming Problems
Available at http://pvamu.edu/aam Applications and Applied Appl. Appl. Math. Mathematics: ISSN: 1932-9466 An International Journal (AAM) Vol. 14, Issue 1 (June 2019), pp. 437 – 451 A New Successive Linearization Approach for Solving Nonlinear Programming Problems 1Inci Albayrak, 2Mustafa Sivri and 3Gizem Temelcan 1,2 Department of Mathematical Engineering Yildiz Technical University Davutpasa, Istanbul, Turkey [email protected]; [email protected] 3 Department of Computer Programming Istanbul Aydin University Kucukcekmece, Istanbul, Turkey [email protected] Abstract In this paper, we focused on general nonlinear programming (NLP) problems having m nonlinear (or linear) algebraic inequality (or equality or mixed) constraints with a nonlinear (or linear) algebraic objective function in n variables. We proposed a new two-phase-successive linearization approach for solving NLP problems. Aim of this proposed approach is to find a solution of the NLP problem, based on optimal solution of linear programming (LP) problems, satisfying the nonlinear constraints oversensitively. This approach leads to novel methods. Numerical examples are given to illustrate the approach. Keywords: Nonlinear programming problems; Taylor series; Linear programming problems; Hessian matrix; Maclaurin series; Linearization approach MSC 2010 No.: 90C05, 41A58, 93B18 1. Introduction Optimization occurs in many fields. Constructing a mathematical model for real life problems is important for optimizers to find optimal strategies effectively. Optimization problems can be classified according to the nature of the objective function and constraints. An optimization problem can be defined as min (or max) of a single (or multi) objective function, subject to (or not to) single (or multi) nonlinear (or linear) inequality (or equality or mixed) constraints. -
Constrained Optimization 5
Constrained Optimization 5 Most problems in structural optimization must be formulated as constrained min- imization problems. In a typical structural design problem the objective function is a fairly simple function of the design variables (e.g., weight), but the design has to satisfy a host of stress, displacement, buckling, and frequency constraints. These constraints are usually complex functions of the design variables available only from an analysis of a finite element model of the structure. This chapter offers a review of methods that are commonly used to solve such constrained problems. The methods described in this chapter are for use when the computational cost of evaluating the objective function and constraints is small or moderate. In these meth- ods the objective function or constraints these are calculated exactly (e.g., by a finite element program) whenever they are required by the optimization algorithm. This approach can require hundreds of evaluations of objective function and constraints, and is not practical for problems where a single evaluation is computationally ex- pensive. For these more expensive problems we go through an intermediate stage of constructing approximations for the objective function and constraints, or at least for the more expensive functions. The optimization is then performed on the approx- imate problem. This approximation process is described in the next chapter. The basic problem that we consider in this chapter is the minimization of a function subject to equality and inequality constraints minimize f(x) such that hi(x) = 0, i = 1, . , ne , (5.1) gj(x) ≥ 0, j = 1, . , ng . The constraints divide the design space into two domains, the feasible domain where the constraints are satisfied, and the infeasible domain where at least one of the constraints is violated. -
Tabu Search for Constraint Satisfaction
Tabu Search for Constraint Solving and Its Applications Jin-Kao Hao LERIA University of Angers 2 Boulevard Lavoisier 49045 Angers Cedex 01 - France 1. Introduction The Constraint Satisfaction Problem (CSP) is a very general problem able to conveniently formulate a wide range of important applications in combinatorial optimization. These include well-known classical problems such as graph k-coloring and satisfiability, as well as many practical applications related to resource assignments, planning and timetabling. The feasibility problem of Integer Programming can also be easily represented by the CSP formalism. As elaborated below, we provide case studies for specific CSP applications, including the important realm of frequency assignment, and establish that tabu search is the best performing approach for these applications. The Constraint Satisfaction Problem is defined by a triplet (X, D, C) where: – X = {x1,· · ·,xn} is a finite set of n variables. – D = {Dx1 ,· · ·,Dxn} is a set of associated domains. Each domain Dxi specifies the finite set of possible values of the variable xi. – C = {C1,· · ·,Cp} is a finite set of p constraints. Each constraint is defined on a set of variables and specifies which combinations of values are compatible for these variables. Given such a triplet, the basic problem consists in finding a complete assignment of the values to the variables such that that all the constraints are satisfied. Such an assignment is called a solution to the given CSP instance. Since the set of all assignments is given by the Cartesian product Dx1 ×· · ·×Dxn, solving a CSP means to determine a particular assignment among a potentially huge search space. -
Constrained Quadratic Optimization: Theory and Application for Wireless Communication Systems
Constrained Quadratic Optimization: Theory and Application for Wireless Communication Systems by Erik Hons A thesis presented to the University of Waterloo in ful¯lment of the thesis requirement for the degree of Master of Applied Science in Electrical and Computer Engineering Waterloo, Ontario, Canada, 2001 c Erik Hons 2001 ° I hereby declare that I am the sole author of this thesis. I authorize the University of Waterloo to lend this thesis to other institutions or individuals for the purpose of scholarly research. I further authorize the University of Waterloo to reproduce this thesis by pho- tocopying or by other means, in total or in part, at the request of other institutions or individuals for the purpose of scholarly research. ii The University of Waterloo requires the signatures of all persons using or pho- tocopying this thesis. Please sign below, and give address and date. iii Abstract The suitability of general constrained quadratic optimization as a modeling and solution tool for wireless communication is discussed. It is found to be appropriate due to the many quadratic entities present in wireless communication models. It is posited that decisions which a®ect the minimum mean squared error at the receiver may achieve improved performance if those decisions are energy constrained. That theory is tested by applying a speci¯c type of constrained quadratic optimization problem to the forward link of a cellular DS-CDMA system. Through simulation it is shown that when channel coding is used, energy constrained methods typi- cally outperform unconstrained methods. Furthermore, a new energy-constrained method is presented which signi¯cantly outperforms a similar published method in several di®erent environments.